What’s Inside?

MIT Solve, in collaboration with researchers from Harvard Business School and the University of Washington (HBS/UW), explored how AI can responsibly augment human judgment in philanthropic decision-making. Facing rising application volumes and limited capacity, Solve sought a way to review submissions efficiently without compromising rigor or fairness.

Solve developed and deployed an AI-powered screening tool for the first round of its Global Challenges process, where application volume is highest and decisions are relatively objective. In 2025, Solve received more than 2,900 applications—a workload that would have required over 190 hours of manual review. The tool cut screening time by 50% while maintaining evaluation quality.

Working closely with academic researchers, Solve also studied AI’s effect on human decision-making. Findings showed that:

• AI assistance improved alignment with expert evaluations.
• Benefits were strongest for novice reviewers, helping close the gap with more experienced evaluators.
• Risks remain around overreliance on AI, particularly when persuasive rationales may influence reviewers to dismiss unconventional but high-potential ideas.

This work demonstrates how AI can improve efficiency and consistency in philanthropy while highlighting the importance of maintaining human oversight.